35
Blue Skies
In fall 1980, Atari, Inc. dispatched the president of its coin-operated games division, Joe Robbins, to Japan to meet with Masaya Nakamura of Namco. Ever since Nakamura made the decision to manufacture Breakout on his own in contravention of his licensing agreement with Atari, relations had been strained between the two firms.1 Robbins, who conducted extensive international business during his time with Empire Distributing and had a solid rapport with Nakamura, was dispatched to smooth things over.
When Namco bought Atari Japan in 1974, it was granted the exclusive right to distribute Atari’s coin-operated games in Japan for a period of ten years.2 After the Breakout situation, Atari began dealing with other companies and granted Sega and Taito the rights to produce upright and tabletop versions, respectively, of Asteroids and Missile Command. When Atari refused to stop licensing to other firms, Namco sued.3 Robbins was sent to Japan to help negotiate a resolution, but he was specifically instructed not to sign anything. Instead, he settled the case.
Robbins’s actions in Japan stunned Atari management, and “who shot JR” jokes began circulating around the company in reference to the infamous cliffhanger that had recently ended a season of the television show Dallas. Combined with the Battlezone manufacturing debacle, signing the settlement with Namco probably cost Robbins his job. Atari ultimately profited from the settlement, however, as it gave the company certain rights to Namco coin-operated games in the home. At the time of the deal, Namco games had been largely unsuccessful in the United States, so this hardly seemed worth the insubordination, but it paved the way for Atari to conclude a deal in April 1981 to acquire the console and home computer rights to the megahit Pac-Man.4 As in the arcade, the arrival of Pac-Man in the home represented the apex of an industry that had experienced rapid growth over the previous year.
***
Just as the coin-operated video game hit a significant new level of popularity in 1980, so too did the home market reach fantastic new heights in 1981. Total revenues for the U.S. market nearly tripled from $475 million to $1.45 billion despite the country being in the throes of a recession.5 Consumers purchased roughly 4.6 million consoles and around 34.5 million cartridges, an average of over seven games per console.6 By the end of the year, 9% of U.S. households that owned a television also owned a programmable video game console.7
Atari remained the dominant console company by far, capturing 65% of the market by dollar volume as it sold an estimated 3.1 million units of the VCS.8 Sales were boosted through a $40 million advertising campaign and by a low price compared to the competition, as retailers began consistently selling the system in the $130–$140 range despite the lack of an official price cut from Atari, choosing to sacrifice hardware margins to capture rapidly expanding software sales.9 In 1980, Atari had achieved fantastic results with its port of the hit Taito arcade game Space Invaders, so for 1981, the company continued to focus on bringing some of the biggest games in the coin-op world into the home. Fortunately for the company, two of the biggest arcade games of 1980, Asteroids and Missile Command, were developed in house. Unfortunately, they both featured additional complexity over Space Invaders, which had already tasked the VCS nearly to its limit. Thankfully, even after the Activision defection, the company still employed two talented programmers in its VCS programming group named Brad Stewart and Rob Fulop.
Brad Stewart graduated with a psychology degree from the University of Pittsburgh in 1972, but had more credits in computer science, for which the school lacked an undergraduate degree at that time. After working for the Ford Motor Company in both Michigan and California for two years, Stewart joined Atari in May 1977 as one of the first hires after the original group that developed the launch titles for the VCS. After working on an unreleased hardware project for a time, Stewart prepared to program his first VCS game just as the programming group decided to port Breakout to the console. Both Stewart and fellow programmer Ian Shepherd wanted the project, so they decided it would go to whichever programmer achieved the highest score on a Breakout cabinet in the break room. Stewart won and coded the game.10
After completing Breakout, Stewart turned his attention to Asteroids. The most significant problem in porting the game to the VCS was the large number of objects that needed to be placed on the screen. Even with all the previously invented tricks like the H-move or venetian blinds, there did not appear to be a credible way to display a dozen or so independently moving objects to recreate the chaotic Asteroids playfield. Stewart’s solution was to “flicker” the graphics, that is, rapidly switch between two sets of graphical elements on every other frame. Stewart discovered that when this technique was executed properly, it was possible to switch between the two sets of graphics quickly enough that it only registered as a subtle flashing of the objects on the screen.11 In this way, Stewart was able to put a large number of rocks up on the screen in a diverse array of colors.
As the game neared completion, one other problem reared its head: the complex collision detection calculations could not be completed within the number of available cycles. This problem could be solved by storing the collision data in tables but doing so would cause the size of the game program to exceed 4K, the maximum amount of memory the VCS hardware could address. Stewart and fellow programmer Bob Smith used every trick they could think of to cram the game into 4K, but they could not do it.12 Fortunately, a group of engineers in the semiconductor group led by Carl Nielsen had developed a method for bank switching on the VCS, a technique that allowed the system to switch between multiple blocks of memory so that even though it is never addressing more than 4K at a time, it can draw from a larger pool of ROM. Bank switching allowed Stewart to program Asteroids in 8K of memory to solve the collision detection problems.13
Asteroids was planned as the big holiday release for 1980, but the complexities of fitting the whole game onto a cartridge resulted in a delay.14 Instead, Atari unveiled Asteroids at the 1981 Winter CES in Las Vegas alongside three other cartridges, Othello, Video Pinball, and Warlords.15 Aside from Asteroids, Warlords was the most interesting game of the set. A two-to-four-player competitive game, it was programmed by Carla Meninsky, a Stanford psychology major who worked with the computers in the school’s neuroscience group to model the human brain. After college she tried to interest companies in letting her build computer animation systems for them and fell in with a group from SAIL with similar ideas. The SAIL people roomed with Atari’s Warren Robinett, who arranged an interview for her with George Simcock at Atari.16
Meninsky was hired by Atari to advance her computer animation work, but with the recent Activision defection, she was asked to write VCS games instead. Given a list of previously brainstormed game ideas, she chose to create a clone of the arcade hit Head On, which Atari released as Dodge ‘Em.17 For her next game, she returned to the list and chose a competitive Breakout concept originally submitted by industrial engineer and graphic designer Roger Hector.18 Warlords allows up to four people to play using the Atari paddle controllers, of which two could be plugged into a single port on the console, and places each in control of a paddle protecting a castle in one corner of the screen made of Breakout-style bricks with a “king” inside. The ball represents a fireball that will destroy any bricks it hits. Players use their paddles to deflect the ball to destroy the other players’ castles and kill the king inside.19 Last man standing wins. Warlords proved the most exciting VCS multi-player experience since Combat back in 1977.
While Stewart toiled on Asteroids, his friend Rob Fulop tackled Missile Command. A native of Oakland, California, Fulop was introduced to computing in 1974 when his alma mater, Skyline High School, installed a terminal hooked up to a computer at the Lawrence Hall of Science. The first program he wrote flipped a coin 100 times and tabulated the results. After that, he wrote a version of Nim and a “Boy Analyzer” designed to entice girls to come to the lab by inviting them to type in the name of a boy at the school and get a printout describing his qualities as a boyfriend.20
Because of his experiences with the terminal at his high school, Fulop attended the University of California, Berkeley and majored in electrical engineering. Fulop had been designing simple games since he was a child, so when he saw an advertisement for a summer programming job at Atari in 1977 during his junior year of college, he applied. Fulop spent the summer in the company’s pinball division, so when he graduated in 1978, he called the company looking for a permanent job. While there were no openings in the coin-op division, he was hired into the Consumer Electronics Division to program games for the VCS.21
Fulop’s first project was a port of the minor Atari coin-op hit Night Driver. He then ported Space Invaders to the Atari 400 and 800, but because he did not want to create a straight up copy of an existing game again, he radically altered the presentation by changing the art for the alien ships, altering their formation, and having them fly in from a giant rocket on the side of the screen. Marketing was nonplussed with the changes, but Atari still shipped the game. After learning of Stewart’s flickering graphics technique while the two were eating lunch together, Fulop decided to employ it to port Missile Command to the VCS. Stung by criticism of his Space Invaders game, he decided not to deviate in any way from the coin-op game and managed to create a version nearly identical to the original.22
Although Asteroids debuted at CES in January 1981, it did not hit retail until August. Missile Command, meanwhile, arrived on store shelves in May. Combined with continuing strong sales of Space Invaders the two games propelled Atari to a dominant 70% share of the cartridge market by dollar volume. Indeed, Space Invaders, Asteroids, and Missile Command accounted for 40% of the cartridges Atari sold that year.23 Overall, the Warner Communications Consumer Electronics Division, which consisted primarily of Atari, surged to $1.2 billion in revenue and income of $287 million as the parent company recorded the best fiscal year in its history with over $3 billion in revenue and income over $226 million.24
Meanwhile, Mattel Electronics was just coming into its own after the advertising campaign featuring George Plimpton turned sales of the Intellivision around at the end of 1980. In 1981, the company spent another $15 million on George Plimpton commercials as it continued to focus on the graphical superiority of the system.25 This led to a bit of controversy as Atari hit back with a commercial of its own featuring a smug boy imitating Plimpton and pointing out that Atari has Asteroids, Missile Command, and Warlords, while Intellivision does not. Mattel countered with a new Plimpton commercial showcasing its own space shooters like Space Battle and two new games called Space Armada and Astrosmash. After this commercial aired, Atari complained to the broadcast standards departments of all three television networks – ABC, CBS, and NBC – and NBC ended up pulling both the Atari and Mattel commercials in early December.26 Despite – or perhaps because of – the controversy, Mattel Electronics more than quadrupled its Intellivision sales to 900,000 units in 1981 and took a 27% console market share by dollar volume.27 Midway through the year, Mattel spun the division out of Mattel Toys so that it reported directly to the president of the entire company.28
In 1979 and 1980, Mattel relied entirely on APh to design and program games for the Intellivision, but with the system proving a success, Mattel Electronics president Josh Denham created an in-house programming group. As at Atari, these programmers were forced to keep their identities a secret to prevent poaching by other companies, but they gained a collective name when TV Guide ran a profile on the Mattel Electronics programmers in its June 19, 1982 issue and dubbed them the “Blue Sky Rangers.”
The first two Blue Sky Rangers, Mike Minkoff and Rick Levine, were already working on handheld games for Mattel Electronics when the group was being formed and transferred to Intellivision. Levine, a former teacher with a math degree from UCLA, became involved in the early hobbyist computer scene and purchased a Sol-20 in 1977. He wrote several games for the computer, including a simulation he used in a psychology class to teach students about different types of stress.29 He subsequently returned to school at University of California Irvine and completed a computer science degree in 1981. While still in school, he learned Mattel was making games and wrote the company asking for a job so he could support himself while earning his degree. He and Minkoff, a 15-year veteran of Mattel who started in data processing before transferring to Mattel Electronics, were sent to APh to learn how the Intellivision worked and start game development. Levine was an avid bowler and had created a bowling handheld for Mattel, so he and Minkoff built on that product to create PBA Bowling as their first game.30
After Intellivision had been on the market for a year, it became clear to management that while the console supported some wonderful sports games, it was suffering in comparison to the VCS because it lacked any decent action games like those currently gobbling quarters in the arcade. Mattel attempted to tackle this problem on several fronts. APh developed a cartridge informally named Some of Theirs that copied some of Atari’s most successful early coin-op games. These included variants of Tank, Jet Fighter, Breakout, and Pong. Mattel’s lawyers requested the Breakout and Pong games be removed, so the final product, called Triple Action, shipped with just the Tank and Jet Fighter games as well as a racing game. The games were named Battle Tank, Biplanes, and Racing Cars, respectively.31 APh also ported Space Invaders to the console as Space Armada. As the Intellivision could only display eight sprites on a line, the programmers used a technique called “sequencing GRAM” that bypassed the exec to directly animate the background graphics to create the invaders.32
The new internal team began developing action games as well. For his second project, Minkoff created a Blockade clone and an adaptation of an unreleased Blockade handheld variant in which the two players control snakes attempting to eat each other’s tail. The games were combined on a single cartridge and released as Snafu. The most successful action game by the company, and indeed the most successful game ever produced for the Intellivision, was Astrosmash, which was created by the first person hired by Mattel specifically to program video games, John Sohl.
A Princeton graduate with a degree in statistics, Sohl moved to California after school to work on a cancer research project at UCLA. After becoming disillusioned by that job and being terminated, Sohl saw an ad in the summer of 1980 for programmers at Mattel Electronics and was hired by Mike Minkoff. Although technically hired to develop handheld product, he was transferred to Intellivision within a matter of weeks. After doing the sounds for Levine and Minkoff’s bowling game, Sohl turned to his first project, a clone of the arcade hit Asteroids. Rather than copying the game exactly, he programmed a variant called Avalanche partially inspired by Missile Command in which the player scored points by shooting rocks falling out of the sky and lost points for any that managed to reach the ground. He also included an enemy called a “spinner” that would kill the player if it reached the ground. When head of product development Richard Chang saw the finished game, he ordered Sohl to go back and make a straightforward Asteroids clone as he had been instructed.33
Sohl developed the Asteroids clone under the name Meteor! but discovered that he had enough space on the cartridge to include Avalanche as well, so he left it in as a second game option. This turned out to be fortunate, because marketing decided that an Asteroids clone would never fly legally and asked him to remove it and just ship his falling rocks game. As other Mattel Electronics staff had become incredibly skilled at the game during this period, Sohl added several more enemy types on higher levels to increase the challenge.34 Mattel released Sohl’s game as Astrosmash in October 1981, and it quickly became the must-have game for the system. In late 1982, it replaced Las Vegas Poker & Blackjack as the game bundled with the system, and by June 1983 it had sold nearly 1 million copies.35 With Astrosmash leading the way, Mattel captured 16% of the software market by dollar volume in 1981.36 For the fiscal year ending January 1982, Mattel reported record sales and earnings of $1.1 billion and $39.1 million, respectively. Mattel Electronics contributed 25% of the parent company’s total sales.37
Magnavox, which in February 1981 ceased to be a subsidiary of Philips and became a brand name for a new entity called North American Philips Consumer Electronics,38 came in a distant third to the two powerhouses on the market, but it still experienced its best year by far, selling around 500,000 consoles to take 7% of the market by dollar share.39 External contractor Ed Averett remained the company’s most significant development asset. In spring 1981, Philips debuted his Asteroids clone UFO! and promoted it heavily. In the summer, the company released a hybrid game by Averett called Quest for the Ring, a tabletop board game for which the Odyssey cartridge is only used to adjudicate combat between the players and a series of monsters encountered while searching a fantasy realm for ten magic rings. Averett created the game in response to letters from parents wishing for a game that introduced strategy alongside reflex-based gameplay to level the playing field when competing with their children.40 In late summer, Philips released a game by Averett that could have easily been the company’s biggest hit and a system seller, but was instead thwarted by the U.S. court system.
During a layover in an airport in February 1981, Averett and Odyssey product manager Mike Staup beheld Pac-Man for the first time. Ed and his wife, Lisa, were busy creating another game at the time, but around May they turned their attention to creating a variant of the coin-op hit. Averett wanted to retain the basic maze chase gameplay but otherwise go in a completely different direction. Staup insisted on something much closer to Pac-Man, but different enough to not risk violating copyright or trademark law, which was unsettled in the area of video games at the time.41 In a compromise, Averett kept the maze, the pursuing enemies, and the dots, but placed only 12 dots in the maze rather than 240. These dots also moved around the playfield over the course of the game. As in Pac-Man, certain of these pellets were special and allowed the player to turn the tables on his pursuers. The protagonist and the pursuers strongly resembled the characters in Pac-Man, though they were slightly different. Mike Staup initially named the character and the game “Munchkin,” but after he joked with Consumer Electronics president Kenneth C. Meinken that they would name the character after him, it ultimately became K.C. Munchkin.42
K.C. Munchkin debuted in August 1981 to brisk sales and favorable comparisons to Pac-Man, which was still unavailable in the home. This caught the ire of Atari, which had recently acquired the home rights to Pac-Man. Both Atari and coin-op rights holder Midway sued in late 1981 and asked for a preliminary injunction to restrain Philips from selling the game until a verdict was reached. The district court judge denied the injunction, but on appeal, the Seventh Circuit Court of Appeals decided that a finding of infringement was likely under the substantial similarity test and therefore granted the injunction in March 1982. The trial dragged on for at least two more years,43 by which time Magnavox’s window to make any money on the game had passed.44
Beyond Atari, Mattel, and Philips, there were essentially no remaining competitors in the North American hardware market. Astrovision, the Columbus, Ohio-based startup that bought the Bally Consumer Division in mid-1980 and continued to market the Bally Arcade, limped along largely by positioning its console as a hobbyist system on which a small, but dedicated group enjoyed using Bally BASIC to create their own games.45 At the January CES, the company showed off yet another version of the long-promised keyboard add-on, this time sporting a visual language invented at the University of Illinois at Chicago called ZGrass,46 but it once again failed to enter production. A handful of new games were released near the end of the year, including a Galaxian clone called Galactic Invasion,47 as was an improved BASIC marketed as Astrovision BASIC,48 but the system did not attract much interest outside its small niche and captured less than 1% of the total market.49
APF fared even worse. At the January 1981 CES, the company debuted an update to its microcomputer hardware called the Imagination System II that integrated the MP-1000 video game system with its computer add-on and incorporated more memory, but it retailed at an unattractive price point of $399 for a system that was limited compared to the competition. An expansion called the System III was also announced that added a disk drive and other peripherals but retailed for an even less attractive $1,195. Lacking faith in this newest offering, the two New York banks that controlled APF’s lines of credit demanded that the company exit the video game and computer businesses immediately and return to its core calculator business or they would call in their loans. Owners Sy and Marty Lipper refused and handed over the keys to the company office at a final meeting. APF essentially ceased operations at that point, though it continued to exist on paper for a few years as the banks wound it down.50
While the hardware market was coalescing around the leaders, the software market began to expand rapidly once Activision proved the viability of third-party software development. Activision itself continued to lead the way in this uncharted field by releasing four more games before the end of the year: Freeway and Kaboom in July and Ice Hockey and Stampede in December. Ice Hockey by Alan Miller was another technical tour de force for the VCS that used several tricks to get four players on the screen and have the two teams wear different color uniforms.51 Stampede by Bob Whitehead started with the programmer’s desire to do a western-themed game because the genre was underrepresented and featured gameplay inspired by the rows of advancing aliens in Space Invaders.52 The player controls a cowboy pursuing cattle that scroll in from the right side of the screen. The player must lasso the steers before they disappear off the left side of the screen. If too may steers escape, the player loses a life.
The most popular Activision games of the year were Freeway by David Crane and Kaboom by Larry Kaplan. The idea for Freeway was born from two events in Chicago during CES. First, Crane and some friends left the convention center through the wrong exit and had to weave through traffic to reach their destination. The same day, Larry Kaplan observed another pedestrian crossing the street against traffic.53 Crane decided avoiding traffic made for an interesting play mechanic and developed a game featuring one or two chickens trying to cross the road across multiple lanes of traffic. Crane pulled off an impressive programming feat by cramming 24 moving sprites onto the screen to represent the traffic and discovered that the gray highway background really made the colorful car graphics pop.54 The fast action and vibrant graphics helped it become a million seller.55
The starting point for Kaboom was an obscure 1978 Atari coin-op game called Avalanche in which several rows of rocks are arrayed at the top of the screen in a similar manner to the bricks in Breakout. These rocks begin to fall a few at a time, and the player must catch them using a set of vertically stacked paddles he can move left and right along the bottom of the screen. Kaplan adapted this basic gameplay, but because he could not display as many objects on the screen at once on the VCS, he greatly reduced the number of falling objects while speeding up their rate of descent to make the game more exciting.56
Kaplan designed the gameplay himself, but for the graphics he turned to Crane, recognized as the best artist of the group. Crane developed a character called the mad bomber that moves back and forth at the top of the screen hurling bombs at the player. Instead of a generic paddle, the player controls a bucket full of water that he whips back and forth to catch all the bombs. Crane added several impressive graphical touches, including flickering fuses on the bombs and little splashes when the bombs land in the bucket.57 The combination of Kaplan’s perfectly balanced twitch gameplay and Crane’s expertly rendered visuals produced yet another million-selling game.58 In all, Activision shipped 4.5 million cartridges in 1981 on its way to sales in the 1982 fiscal year of $65.9 million and an estimated 20% share of the video game market.59
Activision’s success in the market caught the attention of Bill Grubb, who resigned from Atari right before CES in January 1981 in large part because he did not like working for Ray Kassar. Grubb thought there would be good money in serving as a manufacturer’s rep, so he asked Atari to give him Hawaii, which did not currently have a distributor. The company countered with the Pacific Northwest, which did not interest him. Instead, he approached Jim Levy to become the Bay-Area sales rep for Activision.60
Meanwhile, Atari programmer Dennis Koble was becoming unhappy. A UC Berkeley graduate with degrees in electrical engineering and computer science, Koble spent four years working for defense contractor Applied Technology before a friend at Atari invited him to join. The fourth programmer hired into the coin-operated games division, Koble programmed the massive hit Sprint 2, Atari’s Blockade clone Dominoes, the Avalanche game that inspired Kaboom, and Atari’s first pinball machine, the Atarians.61 In 1978, he was drafted by Steve Bristow into Atari’s new handheld game division and created a handheld adaptation of Simon released under the name Touch Me in 1979.62 Once he finished that project, Koble replaced the retiring George Simcock as manager of the VCS programming group at the beginning of 1980. Unhappy as a manager and disillusioned with the layers of bureaucracy at the rapidly expanding Atari, Koble asked management to let him set up a small subsidiary so he could focus on making games without the immense pressure of feeding the VCS beast.63 Koble approached Grubb for advice on how to proceed, and the marketing executive, who missed working for a product development company and was impressed at how much money Activision was making, suggested they acquire some venture funding and become a third-party publisher instead.64
Meanwhile, a junior marketing executive at Mattel Electronics named Jim Goldberger was having similar thoughts about entering third-party development in conjunction with his roommate, Mattel Electronics programmer Brian Dougherty. Grubb had been impressed with Goldberger for some time and had even tried to hire him into Atari, so on April 4, 1981, Grubb, Koble, Goldberger, and Daugherty met at a restaurant and decided to go into business together to provide product for both the VCS and Intellivision.65 Bill Grubb wrote the business plan and presented it to venture capital firm Kleiner Perkins. The organization agreed to come on board with a $1 million investment and introduced Grubb to Steve Merrill at the VC firm Merrill Packard, which committed $500,000. Another half million came from an Indian angel investor that Bill Grubb met through an attorney. In total Grubb secured $2 million in financing to establish the company, which incorporated on June 1, 1981 under the name Imagic.66
In addition to the four founders, Imagic poached several other Atari and Mattel employees. Grubb brought in his friend Mark Bradlee, who had been with him at Black & Decker and was the national accounts manager at Atari, to run sales. Programmer Rob Fulop knew Koble was leaving and asked to come along after being completely nonplussed that his multi-million-selling Missile Command netted him a paltry Christmas bonus of a free turkey from Safeway.67 Bob Smith also defected from Atari, while a programmer named Dave Durran joined his fellows from Mattel. Asteroids programmer Brad Stewart was not there from the start, but he joined in September 1981.68 While the company did not have any products ready to go for 1981, it promised to be a force to rival Activision in 1982.
***
The defection of so many talented programmers to Activision and Imagic between 1979 and 1981 severely impacted the VCS programming group at Atari. These programmers were not only some of the most proficient at squeezing every last ounce of ability out of the system, but they were also the most experienced game developers. Without Alan Miller or David Crane or Brad Stewart around, the new cohort of programmers expected to lead Atari game development into the future were bereft of mentors to guide them through the most efficient ways to harness the capabilities of a system that was extremely difficult to program well. These experienced programmers were missed all the more as Atari continued to license the latest arcade hits to play on a system intended for nothing more complicated than Pong or Tank. Porting the big hits of 1981 to the VCS would require every trick in the book, and the less experienced members of the programming team were not necessarily up to the task.
Perhaps no conversion exemplifies these problems better than the big game of 1982, Pac-Man, programmed by Tod Frye. Frye wrote his first programs in FORTRAN in 1969 at his junior high school, but the institution did not have access to a computer, so the teacher merely informed the students whether their programs would have worked or not. In high school, Frye programmed on an actual time-shared minicomputer for the first time.69 A poor student who preferred smoking pot to attending class, Frye dropped out in his junior year, and his father kicked him out of the house. Frye was homeless for a time, but he eventually landed a job on a construction crew and applied himself to become a master carpenter. Through all these hardships, he never lost his love of programming, and his high school friend Ian Shepard helped him get a job at Atari in 1979.70 After porting Asteroids to the 400 and 800 and working on some handheld projects, Frye was assigned to bring Pac-Man to the 2600.71
Porting Pac-Man with its 240 pellets and five characters to a machine not intended to support more than five sprites was one of the most challenging tasks a VCS programmer was ever asked to carry out. Because of the limited number of sprite objects, the pellets had to be drawn using playfield graphics, which could only be represented as rectangles. As a result, Atari referred to these as “wafers” rather than “pellets” as in the arcade original. Using backgrounds created its own problems, because the VCS only draws the background on one half of the screen and then doubles or mirrors that image to create the other side. This process requires symmetry in the background, but as Pac-Man is eating dots, the background needs to change in an asymmetrical way. Frye used intricate timing techniques to allow both sides of the screen to be updated right before they are drawn, but the cost in limited system resources was immense.72
While getting the gameplay of Pac-Man to work on the VCS was an extraordinary programming feat, the final product was let down by Frye’s inexperience translating coin-operated games to the home. When examining Pac-Man, he felt the key attributes of the game were to collect the dots in the maze and the ability for two people to play together by alternating between each lost life. The aesthetics of the game were less important to him. While the original featured a muted blue maze on a black background so that the bright character graphics would shine through, Frye developed a brown maze on a blue background. The main reason for avoiding black was that VCS programmers were discouraged from using black backgrounds outside of space games because it left CRT televisions more susceptible to burn in, but the result was an ugly playfield that did not look much like its inspiration.73
The decision to accommodate two players created even worse problems. As in Asteroids, showing all the objects on the screen at the same time required the use of a flickering technique to rapidly alternate between two sets of characters. In Asteroids, this process works reasonably well due to the way Brad Stewart balanced the use of the system’s paltry 128 bytes of RAM. Frye was already memory constrained by the need to update the background graphics asymmetrically, but he placed an additional burden on the memory by requiring the game to store the game state and score of a second player so it could be recalled when his turn on the machine began. This did not leave enough RAM to smooth out the flickering on the graphics.74 While Pac-Man himself remains solid, the ghosts flicker rapidly as they pursue the player. Asteroids had noticeable flicker, but because the objects in the game were large rocks moving in set patterns, it was not terribly distracting. When applied to the rapidly moving ghosts of Pac-Man the effect became tiresome on the eyes and made the ghosts harder to track.75 While there was probably no way to avoid some flicker, with more memory Frye could have limited its effects.
Despite its limitations, Pac-Man was still Pac-Man, the hottest game to ever hit the amusement arcade with a starring character that had become the face of the entire burgeoning video game industry. By the end of 1982, Pac-Man had infiltrated all walks of American life. A strategy guide for the game called Mastering Pac-Man cracked the New York Times bestseller list. A novelty duo called Buckner & Garcia released a goofy song called “Pac-Man Fever” in December 1981 that peaked at #9 on the Billboard Hot 100 in March 1982.76 Fleer released a Pac-Man sticker set and sold 100 million packs in less than a year, General Mills produced Pac-Man cereal, and Chef Boyardee added Pac-Man shapes to its canned pasta.77 There were pajamas, lunch boxes, bumper stickers, Hallmark gift cards, and a line of children’s clothing from J.C. Penny emblazoned with the characters from the game.78 In fall 1982, Pac-Man even became the star of his own Saturday morning cartoon produced by Hanna-Barbera. In this environment, anticipation for Atari’s VCS port ran high.
Atari began counting down the days to the Pac-Man release through newspaper ads in late March with slogans like “Soon Pac-Man Will Be Ready for Home Consumption” and “Guess Who’s Coming to Dinner?” This campaign culminated in “National Pac-Man Day” on Saturday, April 3, 1982, when the company launched a $1.5 million PR and marketing blitz in 26 U.S. cities.79 In Chicago, the character appeared on Bozo’s Circus in the morning and at a Cubs-White Sox charity baseball game in the afternoon. In St. Louis, Atari partnered with a local retail chain to host a ten-kilometer run at the conclusion of which Pac-Man was awarded the key to the city. The character showed up at the Dade County Youth Fair in Florida and the Cherry Blossom Parade, in Washington, DC. The day also served as the launching point for a television and radio advertising campaign focused solely on the game.80
Anticipating high demand, Atari manufactured over 1 million cartridges in March, more copies than its top-selling Space Invaders cartridge sold within its first year. Copies of Pac-Man began trickling out to the public in the second half of the month at a suggested retail price of $37.95 and vanished nearly as quickly as they reached store shelves.81 Thanks largely to Pac-Man, Warner Communications reported record first quarter earnings as net income leapt from $49.5 million the year before to $77.9 million. Most of that rise came from a tripling of the profits of the Consumer Electronics Division to $100.6 million. The division, which derived its profits almost entirely from Atari, accounted for two-thirds of Warner’s operating earnings.82 At one of the largest entertainment conglomerates in the world, video games were now a more crucial part of the business than music or movies. The video game appeared to have finally come of age. It would not last.